In expanding into GN writing, I started small by adapting a 14-page screenplay short into GN script format.
In contrast to screenplays, GN stories are broken down per-page instead of per-scene. Shots and angles map into page panels, and narration and voice-overs into captions. Sound effects (SFX) are formatted as dialogue, and actual dialogue may be spoken by characters in-panel or out-panel (O.P.). GN scripts are thus akin to movie shooting scripts.
First off is what happened to the page count: a 14pp screenplay swelled into a GN of first 28pp then 48pp. Quite the expansion!
More intriguing were the ratios of panels/page compared to dialogues/page. Turns out that, for the 15-page samplings of published GNs that I analyzed, the ratios are about equal, with large variances in panel and dialogue count among pages
Sandman vol 1: 4.7±3.6 panels/page (78% variance) vs. 4.5±6.3 dialogues/page.
Sandman vol 8: 6.2±2.1 panels/page (34% variance) vs. 6.5±6.8 dialogues/page;
Note that, over the series, pages become visually more crowded and less varied in complexity, while characters become more talkative overall.
Elfquest book 1: 5.1±2.4 panels/page (48% variance) vs. 5.0±5.7 dialogues/page
Elfquest book 8: 5.1±2.4 panels/page (47% variance) vs. 8.1±4.8 dialogues/page
Note that, over the series, page visual complexity stays exactly the same, while characters become borderline prolix.
Bone Ch7: 5.2±1.7 panels/page (33% variance) vs. 6.1±6.5 dialogues/page.
Bone mostly uses a 2x3-panel format, with relatively few exceptions, hence the lower panels/page variance.
The characters like to talk a lot, too, but sometimes not at all on a given page.
I analyzed several other GNs, with much the same results.
Okay, thought I, time to plug my 28pp GN draft into the same rigorous analysis. The results were eye-opening.
Thingamajigs v 011: 6.4±1.2 panels/page (19% variance) vs. 4.4±2.9 dialogues/page.
This told me that the draft provided little panel variation page-to-page, with characters delivering similar quantities of dialogue/page.
Story aside, such lack of visual and dramatic texture = BORING.
After studying from the masters about how panels are apportioned according to drama and action, I tried again, with a net page count upped to 48.
Thingamajigs v 018: 4.2±2.3 panels/page (54% variance) vs. 3.0±3.3 dialogues/page.
Much improved! Pages and dialogue vary much more. My characters ended up talking/singing more, albeit less on average per page, but the variance is greater and therefore better.
Also, marketability: a 48-page GN is more likely to be picked up than a 28-pager.
Anyway, studying how "the pros" write and illustrate GN is worthwhile, even if one does not use a spreadsheet and statistics to evaluate how one's own works stack up.